public void SimulateEpisode(Random random, int curEpisode, bool shouldExploit)
{
TyTaskNode nodeToExplore = null;
//exploiting:
if (shouldExploit)
{
_sortedNodes.Sort((x, y) => y.TotalValue.CompareTo(x.TotalValue));
//exploit only 50% best nodes:
int count = ((int)(_sortedNodes.Count * 0.5 + 0.5));
nodeToExplore = _sortedNodes.GetUniformRandom(random, count);
}
//explore:
else
{
nodeToExplore = _explorableNodes[curEpisode % _explorableNodes.Count];
}
//should not be possible:
if (nodeToExplore == null)
{
return;
}
PlayerTask task = nodeToExplore.Task;
TySimResult result = TyStateUtility.GetSimulatedGame(_rootGame, task, _analyzer);
nodeToExplore.Explore(result, random);
}
public void InitTree(TyStateAnalyzer analyzer, POGame root, List <PlayerTask> options)
{
_sortedNodes.Clear();
_explorableNodes.Clear();
_nodesToEstimate.Clear();
_analyzer = analyzer;
_rootGame = root;
//var initialResults = TyStateUtility.GetSimulatedGames(root, options, _analyzer);
for (int i = 0; i < options.Count; i++)
{
PlayerTask task = options[i];
var node = new TyTaskNode(this, _analyzer, task, 0.0f);
//end turn is pretty straight forward, should not really be looked at later in the simulations, just simulate once and keep the value:
if (task.PlayerTaskType == PlayerTaskType.END_TURN)
{
TySimResult sim = TyStateUtility.GetSimulatedGame(root, task, _analyzer);
node.AddValue(sim.value);
}
else
{
_explorableNodes.Add(node);
_sortedNodes.Add(node);
}
_nodesToEstimate.Add(task, node);
}
}
public void Explore(TySimResult simResult, System.Random random)
{
if (simResult.IsBuggy)
{
AddValue(simResult.value);
return;
}
POGame game = simResult.state;
List <PlayerTask> options = game._game.CurrentPlayer.Options();
PlayerTask task = options.GetUniformRandom(random);
TySimResult childState = TyStateUtility.GetSimulatedGame(game, task, _analyzer);
if (childState.task.PlayerTaskType != PlayerTaskType.END_TURN)
{
Explore(childState, random);
}
else
{
AddValue(simResult.value);
}
}
